Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Learning to play Go using recursive neural networks.

Lin Wu1, Pierre Baldi

  • 1School of Information and Computer Sciences, Institute for Genomics and Bioinformatics, University of California Irvine, Irvine, CA 92697, USA.

Neural Networks : the Official Journal of the International Neural Network Society
|April 19, 2008
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Heavy-duty diesel vehicles dominate vehicle emissions in a tunnel study in northern China.

The Science of the total environment·2018
Same author

Correction to: SALP, a new single-stranded DNA library preparation method especially useful for the high-throughput characterization of chromatin openness states.

BMC genomics·2018
Same author

Combined effect of apigenin and ferulic acid on frozen-thawed boar sperm quality.

Animal science journal = Nihon chikusan Gakkaiho·2018
Same author

Economic Burden of Individual Suffering from Atrial Fibrillation-Related Stroke in China.

Value in health regional issues·2018
Same author

Assessment of calcium sulfate hemihydrate-Tricalcium silicate composite for bone healing in a rabbit femoral condyle model.

Materials science & engineering. C, Materials for biological applications·2018
Same author

Beyond Low-Rank Representations: Orthogonal clustering basis reconstruction with optimized graph structure for multi-view spectral clustering.

Neural networks : the official journal of the International Neural Network Society·2018
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

IGFD-Net: Illumination-guided frequency decoupling for polarization image fusion.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Multiple-Strategies dung beetle optimizer and its applications in engineering optimization and bankruptcy prediction.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Aggregating global-scale pixel-wise forgery cues within a graph.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Finite-Time intermittent control for secure synchronization of Neutral-Type stochastic delayed neural networks under aperiodic DoS attacks.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces a novel machine learning approach for the game of Go, developing a scalable evaluation function using recursive neural networks. The system demonstrates effective learning from amateur data, transferring skills to professional game levels.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Game Theory

Background:

  • The game of Go presents significant challenges for AI development.
  • Existing computer Go programs do not match skilled human player performance.
  • Large game datasets enable new machine learning approaches.

Purpose of the Study:

  • To develop a scalable machine learning approach for learning Go evaluation functions.
  • To address the challenge of integrating local tactical patterns into a global strategy.
  • To enable AI to learn and play Go at a competitive level.

Main Methods:

  • Developed a machine learning system to automatically learn local pattern propensities from game data.
  • Utilized recursive neural networks, based on probabilistic Bayesian networks, to integrate local information.

Related Experiment Videos

  • Trained the system using local area targets derived from human player game datasets.

    • Main Results:

    • The system effectively integrates local tactical information across the board using recursive neural networks.
    • The aggregated probabilities provide a strategic evaluation function estimating expected game area.
    • Skills learned on smaller boards (N^2) transferred effectively to larger boards (N^2).

    Conclusions:

    • A machine learning approach can learn a scalable evaluation function for Go.
    • AI trained on amateur 9x9 games shows surprising proficiency on professional 19x19 games.
    • This method offers a promising direction for advancing AI in complex board games.